Relationship between biophysical parameters and synthetic indices derived from hyperspectral field data in a coastal marsh from Buenos Aires Province, Argentina

Abstract

ABSTRACT. Coastal wetlands provide essential ecosystem services widely recognized, wich functions and values depend on their ecological integrity. Remote sensing provide environmental indicators of coastal wetlands at landscape scale, but a a deeper understanding is needed on the interactions between field biophysical parameters and the electromagnetic signal. The goal of this work is to analyze and quantify the influence on the spectral response of biophysical parameters -including biomass and leaf area index (LAI)- characterizing stands of Spartina densiflora marshes in Mar Chiquita coastal lagoon, Argentina on their spectral response. Spectral reflectance at high resolution was measured in S. densiflora canopies under natural conditions, manipulating standing biomass by means of successive harvestings. Reflectance data were acquired using a sprectroradiometer in visible, near infrared (IR) and shortwave IR bands. Spectral Vegetation Indices (VI) were calculated for each biomass-LAI situation. Several VI significantly correlated with biomass and LAI, including those 1) based on the red-IR edge (IRI: IR Index, 695/760 ratio, SR: Simple Ratio, NDVI: Normalized Difference VI), 2) based on the sharp change green -IR (GNDVI: green NDVI, 800/550 ratio) and 3) those with a correction for soil noise (OSAVI: Optimized Soil Adjusted VI, MSAVI: Modified SAVI). The indices that showed significant regressions with biomass and LAI were IRI, NDVIAmbar and REIP. The total and green biomass showed better adjustments than LAI, showing R2 values of 0.5. These values were obtained with REIP index. Results indicate that LAI and biomass of Spartina densiflora stands could be determined from spectral data but attention would be payed if large LAI values are found due to the logaritmic adjustments obtained.